On 5 May 2015 at 09:53, Ian Romanick <[email protected]> wrote: > On 04/29/2015 06:14 PM, Dave Airlie wrote: >> From: Dave Airlie <[email protected]> >> >> instead of doing the attempts at dual slot handling here, >> let the backend do it. > > Just curious... why? It seems like everyone wants the same mapping, so > doing it one place ought to be better. There is a lot of stuff in the > GL specs to anticipate hardware that has never materialized.
Primarily to have the mesa core and glsl core bits be consistent, since we need for all the query interfaces etc to work as per the GL API which is double/dvec2/dvec3/dvec4 take one attribute slot at the GL API, So keeping all the mesa/glsl code based around that reasoning, and lowering in the backends seems less likely to have corner cases bugs in things like Get/Bind areas. Dave. _______________________________________________ mesa-dev mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/mesa-dev
